Project Story

Operating in the energy market since 2012, YEPAŞ is the electricity supplier of approximately 5 million people within the scope of nearly 2.2 million subscription agreements in wide geography covering the provinces of Samsun, Ordu, Çorum, Amasya, and Sinop (Yeşilırmak Distribution Region). YEPAŞ provides electricity supply services to a wide range of customers, from residential customers to high-consumption industrial establishments to large-scale business subscribers.

Current Structural Weaknesses and Project Requirements

At the beginning of the project, YEPAŞ, Çalık Holding's subsidiary in the energy market, was managing regular measurement and billing processes for its 2.2 million subscribers. The company collected electricity production data from electricity distribution plants and transferred them to a central data lake, also added current consumption statistics to these data, predicting future electricity consumption habits and trying to plan production capacity according to this statistical estimation in routine workflows. In this process, they managed compelling processes such as data collection, platform maintenance - management, monitoring, and backup through distributed structures, while procuring data collections from multiple sources, he had to deal with data multiplexing handicaps during the orchestration of this scattered data. Yepaş needed a powerful solution to strengthen business continuity and to collect, monitor and back up all data flows in a single point, in short, to manage them with high efficiency by establishing an integrated data warehouse structure regardless of the database format (SQL - noSQL).

The highlights as a result of the Vulnerability Analysis are:

• All business departments were collecting only the data they needed and needed a common repository solution among employees. The collection of different data by different business units in different ways caused data multiplexing problems. These dispersed, derivative, and repetitive datasets had to be deduplicated, and inefficient business processes had to be eliminated. It was critical to create a business department that owns and manages data in order to create common data repositories. Due to the distributed databases in use, it was not possible to collect data with a more fluent and fast method.

Solution Plan and Implementation

Within the scope of the project, the transition phase, which includes the creation of the data lake Yepaş needs with the Bulutistan DBaaS (Database as a Service) solution and the transfer of 200 million data to the new data lake in the first place, was completed with a 100% success rate at the end of the 3-month productive project period.

• PoC tests were defined and applied. Performance and stress tests were carried out. After the PoC approval, 1 TB of data was injected in the first stage. 200 million records were imported with ETL job.

Advantages Provided

The impact report of the YEPAŞ project, which transitioned to the Bulutistan infrastructure, the pioneer of the DBaaS market, as a result of an efficient 3-month project process is given below. Business Continuity Impact

Due to the difficulties experienced during the distributed collection of data, the problem of not being able to make real-time predictions and having to make predictions from past time data has been eliminated. Administrative services related to Data Collection were installed by Bulutistan. In this way, more efficient use of internal resources was ensured. It is planned to solve all problems that may occur at the DBaaS platform level during data collection or processing by Bulutistan.

Performance and Cost Impact

• The entire large data center can be aggregated into a data repository (Data Lake). The loss of efficiency caused by data multiplexing has been prevented. Big data started to be collected in a more efficient way. In this way, it provided time savings of up to 22% in intensive data flows. By collecting less data in fewer databases, a 20% to 60% cost advantage was achieved in terms of hardware and labor resources. Forecasting accuracy in Predictive Analysis processes has been improved by 62%.

Functional Achievements

• The database topology has become dynamically viewable. Analyzes related to queries on the database can be viewed on visual screens. Database backups and backup plans can be created. Database-related user and RBAC roles can be defined. Databases can be monitored and alarms can be generated. Database nodes have a decomposable structure as Read and Write. It paved the way for the creation of custom Dashboards.